Arnold pointed out that Kirk's limiting is working (thanks Arnold).
In response to Wayne's question, no.
My server receives on average 15 req/second, this means that each second
it is possible 15 new hosts might exist leaving 585 historic hosts, in 2
seconds that would be 570 historic hosts, 15 1 second old hosts, and 15
new hosts.
You can calculate based on 15 *unique* req/second how many seconds worth
of history can be stored in 600 entires.
600 / 15 *unique* = 40 seconds.
If you have 1 client sending a request every 2 seconds then you have on
average 14.5 *unique* req/second so you can once again calculate how many
seconds worth of history can be stored in 600 entires.
600 / 14.5 *unique* = 41 seconds.
In answer to one of Arnold's questions:
0x100 = notrap
0x400 = Demobilize AKA KOD
--
.-------------------------------------.
( Biggest security gap -- an open mouth )
`-------------------------------------'
--
Paul-Andrew Joseph Miseiko
On Thu, 22 Sep 2005, wayne wrote:
In <[EMAIL PROTECTED]> Kirk Strauser <[EMAIL PROTECTED]> writes:
Out of the 600 entries in my monlist right now, not one single row has the
0x020 bit of the code set. They're all 0x590 or 0x5b0.
[...]
It looks like a problem with my ntpd, then, doesn't it? I literally never
seem to limit a single client, even the many that are pestering me once
ever second or two seconds. Any ideas?
Could this be related to the problem of ntpd keeping track of only 600
clients while the typical pool server has at least twice that many
regular clients and maybe 4-5 times as many when the server is in the
pool?
-wayne
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
